mmcv
mmcv copied to clipboard
Support receptive field search of CNN models. (TPAMI paper rf-next)
Motivation
Merging a general receptive field search method to mmcv and mmdet. (paper: RF-Next: Efficient Receptive Field Search for Convolutional Neural Networks TPAMI 2022 pdf)
Modification
The RFSearch module is included.
BC-breaking (Optional)
Does the modification introduce changes that break the backward-compatibility of the downstream repositories? If so, please describe how it breaks the compatibility and how the downstream projects should modify their code to keep compatibility with this PR.
Use cases (Optional)
If this PR introduces a new feature, it is better to list some use cases here, and update the documentation.
This pr is a part of pr in mmdet https://github.com/open-mmlab/mmdetection/pull/8191
Checklist
Before PR:
- [x] I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
- [x] Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
- [ ] Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
- [ ] New functionalities are covered by complete unit tests. If not, please add more unit test to ensure the correctness.
- [ ] The documentation has been modified accordingly, including docstring or example tutorials.
After PR:
- [x] If the modification has potential influence on downstream or other related projects, this PR should be tested with some of those projects, like MMDet or MMCls.
- [x] CLA has been signed and all committers have signed the CLA in this PR.
Hi @gasvn , thanks for your contributions. Please add docstring and type hints for your added files.
Hi @gasvn , thanks for your contributions. Please add docstring and type hints for your added files.
I have added docstring and type hints.
Is there any further update about this pull request?
Is there anything I can do to make some progress? Thanks~
Is there anything I can do to make some progress? Thanks~
Hi @gasvn , sorry for my late reply. I'm thinking about where to put the implementation of this hook.
Hi, I have solved the comments above. Please let me know if there is anything I can do. Thanks.
Hi, I have solved the comments above. Please let me know if there is anything I can do. Thanks.
Hi @gasvn , sorry for my late reply.
Hi @gasvn , is there any progress?
Hi @gasvn , is there any progress? Sorry for the delay. We have fixed the issues you mentioned. And we add the support to asymmetric kernel size (1 x n) to support more networks.
@zhouzaida It seems the old pre-commit check we used is out of data, so the revised code cannot pass the lint check. Do we need to pull the latest update from mmcv? Or we just keep it as it is?
@zhouzaida It seems the old pre-commit check we used is out of data, so the revised code cannot pass the lint check. Do we need to pull the latest update from mmcv? Or we just keep it as it is?
It seems the lint passed.
@zhouzaida It seems the old pre-commit check we used is out of data, so the revised code cannot pass the lint check. Do we need to pull the latest update from mmcv? Or we just keep it as it is?
It seems the lint passed.
Thanks, we have solved it.
@zhouzaida This pull request has been processed for several months. Is there a rough schedule for finishing this? Thanks~
@zhouzaida This pull request has been processed for several months. Is there a rough schedule for finishing this? Thanks~
Hi @gasvn , sorry for my late reply. Thanks for your contribution. This PR will be merged this month.
Could we add some unit tests to cover those modifications?
Hello, we have fixed the issues following your suggestions. Thanks. @zhouzaida
Could we add some unit tests to cover those modifications?
It is nice to add some unit tests. But I wonder who should do this job, me or you? If we should do this job, is there any guidance to complete this job. @zhouzaida
Could we add some unit tests to cover those modifications?
It is nice to add some unit tests. But I wonder who should do this job, me or you? If we should do this job, is there any guidance to complete this job. @zhouzaida
Hi @lzyhha , it is best for you to add if you have free time. https://github.com/open-mmlab/mmcv/tree/master/tests/test_cnn and https://github.com/open-mmlab/mmcv/blob/master/tests/test_runner/test_hooks.py would be helpful.
Ok, I'll add the corresponding unit tests within a few days.
Above issues have been fixed.
Are there any other issues we should do apart from unit tests? @zhouzaida
Hello, we have added unit tests for the rfsearch. Please let me know if there are any issues or anything things I can do. Thanks. @zhouzaida
All above issues are fixed.
Please move the file tests/test_cnn/test_rfsearch_operator.py
to tests/test_cnn/test_rfsearch/test_operator.py
and those modifications in tests/test_runner/test_hooks.py
to tests/test_cnn/test_rfsearch/test_search.py
Please move the file
tests/test_cnn/test_rfsearch_operator.py
totests/test_cnn/test_rfsearch/test_operator.py
and those modifications intests/test_runner/test_hooks.py
totests/test_cnn/test_rfsearch/test_search.py
I have completed it.
Hi @lzyhha , have you tested this latest commit with https://github.com/open-mmlab/mmdetection/pull/8191?
Hi @lzyhha , have you tested this latest commit with open-mmlab/mmdetection#8191?
I will test with official mmdetection and our RF-mmdetection as soon as possible.
RF-mmdetection need some corresponding changes, which will be updated later.
Hello, it seems that there is an error in build_cu102 (Error: retrieving gpg key timed out.). Is it a package that failed to download? @zhouzaida
I have tested this latest commit with RF-mmdetection, and everything is OK. And the RF-mmdetection has been updated following the changes in this latest commit.
And what is the schedule to merge this PR?
@zhouzaida